Multi-station time difference positioning method, device and system, and storage medium

Abstract

The invention relates to the technical field of time difference positioning, and discloses a multi-station time difference positioning method, device and system, and a storage medium, the method comprises the following steps: S1, space division: discretizing and dividing a reconnaissance area into grids, and regarding each grid point as a target position; s2, time difference value calculation: calculating time difference values of each grid point relative to a plurality of signal receiving stations, forming a group of time difference data, and storing the time difference data in a time difference value table; and S3, searching: after the reconnaissance station obtains the target time difference value, searching the time difference value closest to the actual measurement time difference value in the time difference value table so as to match the relationship between the actual measurement time difference value and the target position, and further obtaining the current position information of the target. According to the invention, the problems of low efficiency of processing large-scale data and the like in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of time difference positioning, and specifically relates to a multi-station time difference positioning method, a storage medium, a device and a system. Background Art

[0002] With the continuous development of technology, the application of time difference positioning technology is becoming more and more extensive. As an effective positioning means, the multi-station time difference positioning method realizes the precise positioning of the target by collecting the time differences of signals received by multiple stations.

[0003] However, the existing multi-station time difference positioning methods still have some challenges in terms of the efficiency of processing large-scale data and improving the positioning accuracy. Therefore, it is of great significance to develop an efficient and accurate multi-station time difference positioning method. Summary of the Invention

[0004] To overcome the deficiencies of the prior art, the present invention provides a multi-station time difference positioning method, a storage medium, a device and a system, which solve the problems such as the low efficiency of processing large-scale data existing in the prior art.

[0005] The technical solution adopted by the present invention to solve the above problems is as follows:

[0006] A multi-station time difference positioning method includes the following steps:

[0007] S1, Spatial division: discretize the reconnaissance area into grids, and regard each grid point as a target position;

[0008] S2, Time difference value calculation: calculate the time difference values of each grid point relative to multiple signal receiving stations, form a set of time difference data and store it in a time difference value table;

[0009] S3, Search: when the reconnaissance station obtains the target time difference value, search in the time difference value table for the time difference value closest to the actually measured time difference value, so as to match the relationship between the actually measured time difference value and the target position, and further obtain the current position information of the target.

[0010] As a preferred technical solution, in step S3, a search is carried out by combining global rough search and local fine search.

[0011] As a preferred technical solution, during global rough search, use the actually measured time difference value to conduct a rough search within the global range, and find the set of grid points that meet the set conditions as the rough positioning points.

[0012] As a preferred technical solution, during local fine search, with the rough positioning points as the center, divide grids with lower density and reduce the step size for further search.

[0013] As a preferred technical solution, the KNN algorithm is used to search for the time difference value closest to the actually measured time difference value in the time difference value table.

[0014] As a preferred technical solution, the KNN algorithm includes the following steps:

[0015] K1, Initialize the training set: Use the time difference data corresponding to the grid points as the training set samples;

[0016] K2, Extract the time difference data of different density grids as time difference samples;

[0017] K3, Calculate the Euclidean distance between the time difference samples and the extracted training set samples;

[0018] K4, Sort the training set samples in ascending order according to the Euclidean distance;

[0019] K5, Select the first K time difference data training samples with the smallest Euclidean distance;

[0020] K6, Return the grids corresponding to these K time difference data training samples, which are the calculated values of the target position.

[0021] As a preferred technical solution, in step K3, the methods for calculating the Euclidean distance include one or more of the following: direct solution method, pseudo-inverse method, method of moving terms to construct a square matrix, and least squares iterative method based on Taylor series expansion, etc.

[0022] A storage medium stores a program for executing the described multi-station time difference positioning method.

[0023] A multi-station time difference positioning device includes the described storage medium.

[0024] A multi-station time difference positioning system for implementing the described multi-station time difference positioning method includes the following modules connected in sequence:

[0025] Spatial division module: used to execute step S1;

[0026] Time difference value calculation module: used to execute step S2;

[0027] Search module: used to execute step S3.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] (1) The present invention can achieve accurate positioning of the target, and the positioning speed is relatively fast, meeting the real-time requirements;

[0030] (2) The present invention has good robustness. Description of the Drawings

[0031] Figure 1 Flow chart of a multi-station time difference positioning method according to the present invention;

[0032] Figure 2 Physical model diagram of multi-station passive time difference positioning;

[0033] Figure 3 Schematic diagram of time difference from grid points to each reconnaissance station;

[0034] Figure 4 Diagram of the relationship between the original track and the position of the station;

[0035] Figure 5 Track information diagram obtained by back-calculation after time difference is found in data search;

[0036] Figure 6 Comparison diagram of the back-calculated target position and the target true value;

[0037] Figure 7 Error curve diagram when the time difference measurement error is 20 ns;

[0038] Figure 8 Error curve diagram when the time difference measurement error is 10 ns to 50 ns. Specific implementation manners

[0039] The present invention will be further described in detail below in conjunction with embodiments and the accompanying drawings, but the implementation manners of the present invention are not limited thereto.

[0040] Embodiment 1

[0041] As Figures 1 to 8 shown, the present invention relates to a multi-station time difference positioning system based on digital search. The system calculates all possible time difference results by calculating the position information of the target area and the receiving stations, and stores them in the computer memory. During actual detection, a fast data search algorithm is used to search a set of obtained time difference results to obtain the corresponding time difference and back-calculate the target position.

[0042] In order to overcome the defects of the prior art, the present invention has three objectives:

[0043] (1) Propose a time difference positioning algorithm for a passive positioning system using the KNN (k-nearest neighbor) algorithm;

[0044] (2) Obtain the corresponding target position by searching the time difference value, reducing the time and cost of the time difference positioning algorithm;

[0045] (3) By dividing the reconnaissance area into grids, the complexity of data processing is reduced and the positioning efficiency is improved.

[0046] The present invention proposes a high-precision time difference positioning algorithm based on the KNN data search strategy. Through spatial division, the reconnaissance area is discretized, each grid point is regarded as a target position, and the time difference values of each grid point relative to multiple signal receiving stations are pre-calculated and a set of time difference data is stored in a table. When the reconnaissance station obtains a set of target time difference values, the KNN algorithm is used to search in the time difference value table for the time difference value closest to the actually measured time difference value, so as to quickly match the relationship between the measured value and the target position, and then obtain the current position information of the target, complete high-precision positioning and form a track. To improve the positioning efficiency and real-time performance, the present invention also proposes a strategy combining global rough search and local fine search. Compared with the traditional time difference positioning algorithm, the algorithm of the present invention shows significant advantages in terms of positioning accuracy and real-time performance, and avoids the complex equation solving process and multiple repeated iteration processes of the traditional time difference positioning method, greatly improving the real-time performance of the system. Through simulation verification, it is proved that the algorithm of the present invention shows significant advantages in terms of positioning accuracy and real-time performance. The algorithm flow chart is as Figure 1 shown.

[0047] The multi-station time difference positioning method based on the data search algorithm provided by the present invention includes the following aspects:

[0048] (1) Preparation of time difference data for the training set:

[0049] The passive detection system is composed of n (n≥3) measured stations, and the multi-station passive positioning physical model is as Figure 2 shown:

[0050] Among them, (x, y, z) T is the spatial position of the radiation source, (x i , y i , z i ) T , i = 0, 1, 2,..., n are the positions of each measured station, i = 0 represents the main station, i = 1, 2,..., n represent the secondary stations, and r i is the distance between the radiation source and the i-th station.

[0051] Then the distance between the measured station i and each station is:

[0052] r i = [(x - x i ) 2 + (y - y i ) 2 + (z - z i ) 2 ) 1 / 2 (1)

[0053] The distance from the target to the i-th station and the distance from the target to the main station (x 0 , y 0 , z 0) T The distance difference Δr i is as follows:

[0054] Δr i = r i - r 0 (2)

[0055] It can be seen from equations (1) and (2) that the passive positioning equation set is:

[0056]

[0057] where Δr i = TDOA i0 × c, c is the speed of light, and TDOA i0 is the signal arrival time difference between the i-th station and the master station.

[0058] The spatial position (x, y, z) of the radiation source can be obtained by solving the equation T . Common solution methods include: direct solution method, pseudo-inverse method, method of moving terms to construct a square matrix, and least squares iterative method based on Taylor series expansion, etc.

[0059] The positioning area is refined into several small grids, and each grid point is assumed to be a possible position of the target. Suppose there are three reconnaissance stations in the target area, and the master station and two slave stations are located at points A, B, and C respectively. The time differences Δt AB , Δt AC from the grid point P to the two slave stations B and C and to the master station A can be calculated by the following formula:

[0060] Δt AB = (PB + AB - AP) / c (4)

[0061] Δt AC = (PC + AC - AP) / c (5)

[0062] By traversing all grid points in the target area, the time difference values corresponding to each grid point can be calculated and the results can be stored in tabular form. The schematic diagram of the time difference from the grid point to each reconnaissance station is as Figure 3 shown.

[0063] (2) Application of KNN algorithm:

[0064] Using the constructed time difference data set, the KNN algorithm is used to process the new measured time difference values. By calculating the similarity between the new time difference values and the existing data in the data set, the nearest K neighbors are found, and the position of the target is estimated based on the labels of these neighbors (i.e., the corresponding grid points). The main algorithm description of the application of the KNN algorithm in passive time difference positioning is as follows:

[0065] 1. Initialize the training set, i.e., the time difference data corresponding to the grid points;

[0066] 2. Extract time difference data of different densities from the grid according to actual needs as time difference samples to improve real-time performance.

[0067] 3. Calculate the Euclidean distance between the test time difference samples and the time difference training set samples corresponding to the extracted grid;

[0068] 4. Sort the training set samples in ascending order according to the Euclidean distance;

[0069] 5. Select the first K time difference data training samples with the smallest Euclidean distance;

[0070] 6. Return the grids corresponding to these K time difference samples, i.e., the calculated values of the target position.

[0071] (3) Search grid division method:

[0072] It should be noted that the performance of the nearest neighbor algorithm is affected by the size and distribution of the time difference value table. Therefore, in practical applications, it is necessary to reasonably design the grid size and quantity, and collect sufficient sample data to construct an accurate time difference value table.

[0073] To improve the positioning efficiency and real-time performance, the present invention proposes a strategy combining global rough search and local fine search. First, use the actually measured time difference values to perform a rough search globally and find a set of grid points that meet certain conditions. The purpose of this step is to quickly narrow the positioning range and reduce the complexity of subsequent calculations. Then, with these roughly located points as the centers, divide fine grids and set a small step size for further search. The division of the fine grids should be adjusted according to actual needs, ensuring both positioning accuracy and avoiding excessive calculation due to being too fine.

[0074] Since in the actual passive reconnaissance process, the target position is relatively slowly changing, after obtaining the initial target position, the grid can be set for the track points formed by this target, and including the next track point in the subdivided grid area can complete the time difference search for the next point, without the need to perform a global rough search for each point.

[0075] By combining global rough search and local fine search, the method of the present invention can effectively reduce the calculation amount and improve the real-time performance while ensuring the positioning accuracy.

[0076] Embodiment 2

[0077] As Figures 1 to 8 shown, based on Embodiment 1, this embodiment provides a more refined implementation manner.

[0078] To verify the effectiveness of the KNN algorithm in passive time difference of arrival (TDOA) location, we conducted a simulation analysis. First, we constructed a signal propagation model that included three receivers and a target. Then, we used the KNN algorithm to process the received signals and calculated the location of the target.

[0079] To further verify the practical application effect of the KNN algorithm in passive TDOA location, we conducted an experimental verification. We built an actual wireless location system and collected a certain amount of experimental data using this system. Then, we used this data to train and test the KNN algorithm. The experimental results show that the KNN algorithm can achieve precise location of the target in practical applications, and the location speed is relatively fast, meeting the real-time requirements.

[0080] First, referring to Figure 1 the detailed method shown, a signal propagation model that included three receivers and a target was constructed according to the operation steps. In the simulation calculation, when considering that the main station and the slave stations form an isosceles triangle with an included angle of 120° and the distance between the main station and the slave stations is 30 km, the differential measurement accuracy is 20 ns, and the k value is taken as 1 in the simulation. It is assumed that the standard deviations of the time difference measurement errors at each station are all equal. The grid is set with a width of 5 km. Calculate the time difference corresponding to the time when the target grid arrives at the main station and the slave stations and store it in the computer. Set a track within the target grid area, use the KNN algorithm to search for the time difference corresponding to the target track. When the search result meets the requirements compared with the stored result, the corresponding grid position is the estimated target position. After traversing all target points, the track result of the target can be obtained. The location accuracy of the target can be obtained by calculating the error between the true track and the estimated track location points.

[0081] In the simulation calculation, different situations such as when the main station and the slave stations form an isosceles triangle with an included angle of 120° and the distance between the main station and the slave stations is 30 km, and the differential measurement accuracy is in the range of (10 ns - 50 ns) are considered simultaneously, and the k value is taken as 1 in the simulation.

[0082] It is assumed that the standard deviations of the time difference measurement errors at each station are all equal. Figures 4 to 7 The process of generating the target track and the comparison between the generated track and the true target track are given when the time difference measurement error is 20 ns in engineering.

[0083] It can be seen that: the back-calculated time difference track is continuous and the location accuracy is relatively high. From Figure 7 It can be seen that when the time difference measurement error is 20 ns, the CEP value of the entire track segment is 0.26 km. Compared with the conventional engineering index of 2%R (8 km at 400 km), the accuracy has been greatly improved. Figure 8The curve of the change of the CEP index of the positioning accuracy with the time difference measurement error is given when the time difference measurement error is 20 ns to 50 ns. It can be seen that as the time difference measurement error increases, the measurement accuracy decreases, which is consistent with engineering experience. However, it should also be noted that the target positioning accuracy can still be maintained within 0.65 km. The simulation results show that the KNN algorithm can achieve a high positioning accuracy in the time difference positioning system and has good robustness.

[0084] As described above, the present invention can be preferably realized.

[0085] All the features disclosed in all the embodiments in this specification, or all the steps in the methods or processes implicitly disclosed, except for the mutually exclusive features and / or steps, can be combined and / or extended and replaced in any way.

[0086] The above is only a preferred embodiment of the present invention, and does not impose any form of limitation on the present invention. Based on the technical essence of the present invention, any simple modification, equivalent replacement and improvement made to the above embodiments within the spirit and principle of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A multi-station time difference positioning method, characterized in that: The following steps are involved: S1, spatial division: discretize the reconnaissance area into grids, and regard each grid point as a target location; S2, time difference calculation: calculate the time difference value of each grid point relative to multiple signal receiving stations, form a set of time difference data and store it in the time difference value table; S3, search: when the reconnaissance station obtains the target time difference value, it searches the time difference value table for the time difference value closest to the actual measured time difference value, thereby matching the relationship between the actual measured time difference value and the target position, and then obtaining the current position information of the target.

2. A multi-station time difference positioning method according to claim 1, characterized in that: In step S3, a search is performed by combining a global rough search with a local fine search.

3. A multi-station time difference positioning method according to claim 2, characterized in that: During the global rough search, the actual measured time difference value is used to perform a rough search in the global range to find a set of grid points that meet the set conditions as the rough positioning points.

4. A multi-station time difference positioning method according to claim 3, characterized in that: When performing local fine search, the coarse positioning point is used as the center to divide the grid into a lower density, and the step size is reduced for further search.

5. A multi-station time difference positioning method according to any one of claims 1 to 4, characterized in that: The KNN algorithm is used to find the time difference value closest to the actual measured time difference value in the time difference value table.

6. A multi-station time difference positioning method according to claim 5, characterized in that: The KNN algorithm consists of the following steps: K1, initialize the training set: take the time difference data corresponding to the grid points as training set samples; K2, extracting time difference data of grids with different densities as time difference samples; K3, calculates the Euclidean distance between the time difference sample and the extracted training set sample; K4, sort the training set samples in ascending order according to the Euclidean distance; K5, select the first K time difference data training samples with the smallest Euclidean distance; K6, returns the grid corresponding to the K time difference data training samples, which is the calculated value of the target position.

7. A multi-station time difference positioning method according to claim 6, characterized in that: In step K3, the method for calculating the Euclidean distance includes one or more of the following: a direct solution method, a pseudo-inverse method, a method of constructing a square matrix by transposing terms, and a least squares iteration method based on Taylor series expansion, etc.

8. A storage medium, characterized in that: A program for executing a multi-station time difference positioning method as described in any one of claims 1 to 7 is stored.

9. A multi-station time difference positioning device, characterized in that: A storage medium comprising the storage medium as claimed in claim 8.

10. A multi-station time difference positioning system, characterized in that: A multi-station time difference positioning method for implementing any one of claims 1 to 7, comprising the following modules connected in sequence: Space division module: used to execute step S1; Time difference calculation module: used to execute step S2; Search module: used to execute step S3.

Images